CN109008936A - Diopter detection method and device based on reference object - Google Patents

Abstract

The invention provides a diopter detection method and device based on a reference object, wherein the method comprises the following steps: presenting a first image on a terminal screen; taking a second image in response to an external operation; identifying a reference object from the second image and determining the image size of the reference object; and determining the diopter of the measured person according to the image size of the reference object. According to the diopter detection method and device based on the reference object, which are provided by the invention, the diopter detection can be finished by a common intelligent terminal without the intervention of professional personnel, and the user can operate and detect the diopter of the user by himself, so that the device and the method have strong convenience.

Description

Diopter detection method and device based on reference object

technical field

The invention relates to the field of medical data analysis, in particular to a method and device for detecting diopter based on a reference object.

Background technique

With the development of modern economy, the advancement of education, and the popularization of various smart devices, the number of myopia is increasing year by year, and it is becoming younger. Therefore, the importance of low-cost and universal diagnostic methods for myopia is self-evident, especially It is for teenagers.

Usually, in eye hospitals and optician centers, optometrists can use professional equipment to judge whether a person's eyeball is myopic and the degree of myopia by measuring the diopter. However, under normal circumstances, people decide to go to a professional institution for optometry when they find that there is a problem with their eyes. Only for myopia. We know that if it is not congenital myopia, teenagers who use their eyes frequently due to fatigue and often use their eyes at close range will first lead to false myopia and further develop into true myopia. Pseudomyopia can be restored by modern medical means. Myopia can only be treated with glasses or surgery. Therefore, if there is a way to allow teenagers to easily detect whether they are myopic, it will be of great significance for the prevention of myopia in teenagers. At the same time, frequent measurement of diopter can also help people who are already myopic to control the development of myopia.

There are three existing diopter detection methods, namely subjective optometry, retinoscopy optometry and computer optometry. Accuracy in descending order. In hospitals or optician centers, computerized optometry is usually used for preliminary examination, and the actual diopter is determined by subjective optometry. The three methods are briefly introduced below.

Subjective optometry uses a series of lenses with different diopters, and with the cooperation of the patient, repeatedly tries lenses with different diopters, so that the patient can restore normal visual acuity through the lens. Subjective optometry needs to be used in conjunction with the eye chart. The patient needs to keep a certain distance from the eye chart. The specific value depends on the eye chart and calculation method, usually six meters, and the light in the room must be comfortable.

The retinoscopy method is to use a retinoscopy mirror to project a beam of light to the patient's eye refractive system to reach the retina, and then the reflected light from the retina reaches the retinoscopy mirror and passes through the retinoscopy mirror peephole (referred to as the retinoscopy hole). observed. This retinal reflected light is the "red light reflex", which is the main basis for retinoscopy analysis. The forward movement and reverse movement formed by the red light reflection are also different according to the different refractive states of the patients. The optometrist analyzes the different motions and removes the corresponding lenses from the standard lens box to eliminate the motions until a neutral point is found. The standard lenses used to find the neutral point are closely related to the patient's refractive status. Retinoscopy is also called retinoscopy, and retinoscopy is also called retinoscopy.

Computerized optometry belongs to objective optometry, and its principle is basically the same as that of retinoscopy. Infrared light source and automatic fog vision device are used to achieve the purpose of relaxing eyeball adjustment, and photoelectric technology and automatic control technology are used to check the diopter.

The common feature of the existing three methods is that they need special equipment and the assistance of professionals. It can be seen that the existing diopter detection method is not convenient.

Contents of the invention

In view of this, the present invention provides a method for detecting diopter based on a reference object, including:

presenting the first image on the terminal screen;

taking a second image in response to an external operation;

identifying a reference object from the second image and determining an image size of the reference object;

The diopter of the subject is determined according to the image size of the reference object.

Preferably, the image size is the image length of the reference object.

Preferably, the diopter of the subject is determined using the following formula:

Wherein, D is the diopter, _h0 is the distance between the reference object and the imaging device when shooting the test image, the test image includes the image of the reference object, and _l0 is the distance of the reference object in the test image Image length, l ₁ is the image length of the reference object in the second image.

Preferably, the image size is the image area of the reference object.

Preferably, the diopter of the subject is determined using the following formula:

Wherein, D is the diopter, _h0 is the distance between the reference object and the imaging device when shooting the test image, the test image includes the image of the reference object, _S0 is the distance of the reference object in the test image Image area, S ₁ is the image area of the reference object in the second image.

Preferably, the identifying the reference object from the second image and determining the image size of the reference object includes:

The second image is used as input data of a machine recognition model, so that the machine recognition model outputs the image size of the reference object, and the machine recognition model is obtained by using a plurality of training images and training labels, wherein the The training image has a portrait and a reference object image, and the training label includes the position and size of the reference object image in the training image.

Preferably, the reference object is a card with a magnetic strip.

Preferably, the identifying the reference object from the second image and determining the image size of the reference object includes:

converting the second image to a grayscale image;

determining the magnetic stripe image of the card according to the lines of the grayscale image;

Determines the size of the magnetic stripe image.

Preferably, before the step of acquiring the second image, it also includes:

Presenting prompt information on the terminal screen to prompt the subject to control the camera on the terminal to take pictures of the subject's face to obtain a second image when the subject can clearly see the first image; and/ Or prompting the detected person to hold the posture of the reference object; and/or prompting a way to adjust the distance between the terminal and the detected person.

Correspondingly, the present invention also provides an electronic device, including: at least one processor and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, The instructions are executed by the at least one processor, so that the at least one processor executes the above reference-based diopter detection method.

According to the diopter detection method and device based on the reference object provided by the present invention, by presenting the image for detecting the diopter on the terminal screen, and the user operates the terminal to capture an image containing the image of the reference object on his own face, and then through the captured image The diopter of the user's eyes can be calculated by the image size of the reference object in the center. This detection process can be completed by ordinary electronic equipment, and does not require the intervention of professionals. It is operated by the user himself, which is very convenient.

Description of drawings

In order to more clearly illustrate the specific implementation of the present invention or the technical solutions in the prior art, the following will briefly introduce the accompanying drawings that need to be used in the specific implementation or description of the prior art. Obviously, the accompanying drawings in the following description The drawings show some implementations of the present invention, and those skilled in the art can obtain other drawings based on these drawings without any creative work.

Fig. 1 is a flow chart of a diopter detection method based on a reference object provided by an embodiment of the present invention;

Fig. 2 is the first image in the embodiment of the present invention;

Fig. 3 is a schematic diagram of a scene where a user uses the reference object-based diopter detection method provided by the present invention;

Figure 4 is the second image in an embodiment of the present invention.

Detailed ways

The technical solutions of the present invention will be clearly and completely described below in conjunction with the accompanying drawings. Apparently, the described embodiments are some of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

In addition, the technical features involved in the different embodiments of the present invention described below may be combined with each other as long as there is no conflict with each other.

An embodiment of the present invention provides a method for detecting diopter based on a reference object. This method can be executed by electronic devices with camera devices such as smart phones and tablet computers, and is used for users to self-test the diopter. As shown in Figure 1, the method includes Follow the steps below:

S1. Present a first image on the terminal screen. The image may be an image specially used in the medical field for checking diopter or astigmatism. The content of the image can make it easy for the user to subjectively judge whether the eyes are focused, such as the image shown in FIG. 2 . Those skilled in the art can understand that the content of the first image is not limited to the content shown in FIG. 2 , and multiple similar images may be stored in the terminal for selection and use by the user during practical application.

The size and resolution of the image should be predetermined according to the parameters of the display screen, so that the user cannot fully see the content of the first image at a long distance, but can clearly see the content of the first image at a short distance. S2, taking a second image in response to an external operation, the user should understand the operation mode when using the solution provided by the present invention, that is, the user needs to cover one eye with a reference object, and observe the first image presented on the screen through the uncovered eye One image, while moving the body or the display device from far to near, so that the first image gradually approaches the eyes, and when the user can see the first image clearly, the user controls the terminal to shoot the second image towards his face. The so-called clear vision means that the pattern on the picture incident on the eyes can just converge on the retina. What the user experiences is that at this distance, the edges in the image are sharp, and even if the user gets closer, the sharpness of the edge will not increase. , this level needs to be judged subjectively by the user, and the second image can be taken when it is determined that this level has been reached.

The second image captured by operating in this manner should include images of the user's face and the reference object. When using a device such as a smart phone or a tablet computer to perform this method, the front camera of the device can be turned on while presenting the first image, and the viewfinder content is not displayed on the screen; of course, an area can also be divided to display the viewfinder content to avoid errors An accurate shooting angle is used to ensure that the second image includes the user's face and the reference object.

The distance between the camera and the user should be the same as the distance between the screen displaying the first image and the user, preferably the camera and the screen are integrated; when implementing this method through separate equipment, it should be ensured that the camera is positioned as close to the display as possible location. There are many ways to operate the external camera to take pictures, for example, it can be a button to shoot, or it can be controlled by voice or facial expression to control the camera and so on. According to the correct method of use, the user should clearly observe the first image on the screen 31 through the eyes not covered by the reference object under the situation shown in FIG. Image, there is a reference object image 41 in FIG. 4 .

S3, identify the reference object from the second image and determine the image size of the reference object, the reference object is a pre-selected object, the object should have a standard size, for example, coins, bank cards, etc. can be selected, the present invention can allow The user uses one of several reference objects. The parameter of image size can have many situations. For different reference objects, the type of size to be obtained is different, for example, it can be the length of a side of the image of the reference object, the overall diameter, the overall area, etc. The reference object used in this embodiment is a standard-sized card (credit card or debit card, etc.), and the length AD of the reference object image 41 in FIG. 4 is selected for subsequent calculation.

S4. Determine the diopter of the subject according to the image size of the reference object. Based on the distance measurement principle of similar triangles, when the user holds the card to take the second image, the closer the card is to the camera, the larger the card image in the image; The farther away, the smaller the card is in the second image. Specifically, within a certain range, through reasonable approximation, the principle of similar triangles can be used to represent the relationship between the image size of the card and the distance between the card and the camera. That is, the distance between the user's eyes and the terminal is determined by the image size of the reference object.

Those skilled in the art can understand that in the medical field, the diopter of the eye is 0, which means that parallel incident light rays can just converge on the retina, which means that objects at infinity can be seen clearly. For normal adults, an emmetropic eye with a diopter of plus or minus 0.75 is an emmetropic eye (no myopia or hyperopia, normal vision). For minors, the diopter range of an emmetropic eye varies with age.

The diopter is equal to the reciprocal of the distance between the object and the human eye when the person just sees the object clearly. For example, a diopter of 0.75 corresponds to 1.33 meters, which means that if an adult sees an object 1.33 meters away very clearly, it means emmetropia. If the object is only 0.5 meters away from the human eye to see clearly (without double image), then the diopter is 2. If you wear myopia correction, the corresponding myopia degree is 200 degrees.

There are many ways to calculate the distance through the image size and then calculate the diopter. For example, you can use the image size of the reference object and the pre-calibrated coefficient to calculate the distance, or use the image size of the reference object and the actual size of the reference object to calculate the distance, and then convert it into diopter value.

According to the diopter detection method based on the reference object provided in the embodiment of the present invention, an image for detecting the diopter is presented on the terminal screen, and the user operates the terminal to capture an image containing the image of the reference object on his own face, and then through the captured image The diopter of the user's eyes can be calculated by the image size of the reference object in the center. This detection process can be completed by ordinary electronic equipment, and does not require the intervention of professionals. It is operated by the user himself, which is very convenient.

In order to guide the user to operate the detection process correctly, some prompt information can also be presented before the above step S2, for example, it can prompt the timing of taking the second image, that is, remind the subject to control the terminal on the condition that the first image can be seen clearly. The camera device shoots the face of the detected person to obtain the second image;

It can also prompt the subject to hold the posture of the reference object, such as holding the reference object to cover one eye;

A way to adjust the distance between the terminal and the detected person may also be prompted, for example, prompting the user to move the terminal from far to near.

Specifically, prompts can be given in the form of text descriptions, animation demonstrations, or a combination of the two.

The display of prompt information can improve the user's operation efficiency and detection accuracy, but it is not necessary to prompt each time before shooting. For example, it can be prompted only when it is used for the first time, or it can be prompted according to the user's choice according to the user's needs. Of course, it is also possible. As help text for users to view at any time.

For the above step S4, the present invention provides two preferred embodiments. In a first preferred embodiment, the image size is the image length of the reference object. Before using this method, two parameters need to be calibrated. Specifically, before using this method, the manufacturer can photograph the reference object at a known distance, then measure the length of the reference object image in the captured image, and call the captured image a test image, and the calibrated parameters are is the shooting distance h ₀ of the test image and the image length l ₀ of the reference object in the test image.

Further, in the above step S4, the following formula can be used to determine the diopter of the subject:

Among them, D is the diopter, h ₀ is the distance between the reference object and the camera device when shooting the test image, the test image includes the image of the reference object, l ₀ is the image length of the reference object in the test image, and l ₁ is the distance between the reference object in the second image The image length of the reference object.

In the second preferred embodiment, the image size is the image area of the reference object, and the image area can be directly recognized from the image by using the number of pixels through the principle of machine vision, or can be obtained by calculation. For example, for a square reference object, the area can be calculated by identifying the length and width of the image; for a circular reference object, the area can be calculated by identifying the diameter. Before using this method, two parameters need to be calibrated. Specifically, before using this method, the manufacturer can photograph the reference object at a known distance, then measure the area of the reference object image in the captured image, and call this captured image a test image, and the calibrated parameters are is the shooting distance h ₀ of the test image and the image area S ₀ of the reference object in the test image.

Further, in the above step S4, the following formula can be used to determine the diopter of the subject:

Among them, D is the diopter, h ₀ is the distance between the reference object and the camera when shooting the test image, the test image includes the image of the reference object, S ₀ is the image area of the reference object in the test image, and S ₁ is the image area of the reference object in the second image. The image area of the reference object.

For the above step S3, the present invention provides two preferred embodiments.

In a first embodiment, a machine learning model can be used to identify the image size. Those skilled in the art can understand that using machine learning requires a large amount of training data to build a model. This embodiment pre-constructs a picture data set containing portraits and standard reference objects, and records the position and size of the standard reference objects in the picture. as a label for the image.

Take a credit or debit card with a magnetic stripe on the back as a standard reference. The picture data set can be obtained by taking pictures of people holding cards covering their eyes, or by computer-generated pictures that simulate actual scenes. Specifically: put cards of different sizes and colors on the portrait picture at any rotation angle and any position to obtain a portrait picture with the card on the face. While placing the card, record the size of the card and its position on the picture.

The recognition model built with image data set and label training enables the model to recognize the position and size of the standard reference object (card) in the image.

The recognition model can be implemented with a convolutional neural network. A large number of image datasets simulating actual scenes and the corresponding card size and position information are used to train the neural network so that it can recognize the size and position of the card in the real scene.

After the recognition model is trained, step S3 is: use the second image as input data of the machine recognition model, so that the machine recognition model outputs the image size of the reference object. The machine recognition model is obtained by using multiple training images and training labels, wherein the training images have portraits and reference object images, and the training labels include the position and size of the reference object images in the training images.

Before providing this product to users, the manufacturer can use various reference objects to train various models, and just introduce the reference objects that can be used to users. Using machine learning to determine the image size of reference objects can be used to a large extent. Improve the flexibility of the program.

In the second embodiment, the traditional machine vision method can be used to determine the size of the image. In this case, there are certain restrictions on the type of reference object, and it is necessary to select a reference object with certain characteristics. In this embodiment, it is preferred to have Card with magnetic stripe.

In this case, step S3 includes the following steps:

S31, converting the second image into a grayscale image, there are many ways to convert the color image into a grayscale image, and it is all feasible to select an existing color conversion method;

S32. Determine the magnetic stripe image of the card according to the lines of the grayscale image. Due to the color characteristics of the magnetic stripe, the image of the magnetic stripe in the grayscale image is very obvious, appearing as a long black strip, which can be found according to the shape characteristics of the lines. the location of the magnetic stripe image;

S33. Determine the size of the magnetic stripe image. Since the magnetic stripe runs through the entire card, the length of the magnetic stripe is the length of the reference object.

Using the traditional machine vision recognition method, the image size of the reference object can be calculated in real time according to the actual situation. This solution has high accuracy.

This embodiment also provides an electronic device, including at least one processor and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions Executed by the at least one processor, so that the at least one processor executes the above reference-based diopter detection method.

Apparently, the above-mentioned embodiments are only examples for clear description, rather than limiting the implementation. For those of ordinary skill in the art, other changes or changes in different forms can be made on the basis of the above description. It is not necessary and impossible to exhaustively list all the implementation manners here. And the obvious changes or changes derived therefrom are still within the scope of protection of the present invention.

Claims (10)

1. A diopter detection method based on a reference object, characterized in that, comprising: presenting the first image on the terminal screen; taking a second image in response to an external operation; identifying a reference object from the second image and determining an image size of the reference object; The diopter of the subject is determined according to the image size of the reference object. 2. The method according to claim 1, wherein the image size is an image length of a reference object. 3. method according to claim 2, is characterized in that, utilizes following formula to determine the diopter of measured person: Wherein, D is the diopter, _h0 is the distance between the reference object and the imaging device when shooting the test image, the test image includes the image of the reference object, and _l0 is the distance of the reference object in the test image Image length, l ₁ is the image length of the reference object in the second image. 4. The method according to claim 1, wherein the image size is an image area of a reference object. 5. method according to claim 4, is characterized in that, utilizes following formula to determine the diopter of measured person: Wherein, D is the diopter, _h0 is the distance between the reference object and the imaging device when shooting the test image, the test image includes the image of the reference object, _S0 is the distance of the reference object in the test image Image area, S ₁ is the image area of the reference object in the second image. 6. The method according to claim 1, wherein the identifying the reference object from the second image and determining the image size of the reference object comprises: The second image is used as input data of a machine recognition model, so that the machine recognition model outputs the image size of the reference object, and the machine recognition model is obtained by using a plurality of training images and training labels, wherein the The training image has a portrait and a reference object image, and the training label includes the position and size of the reference object image in the training image. 7. The method according to claim 1, wherein the reference object is a card with a magnetic strip. 8. The method according to claim 7, wherein the identifying the reference object from the second image and determining the image size of the reference object comprises: converting the second image to a grayscale image; determining the magnetic stripe image of the card according to the lines of the grayscale image; Determines the size of the magnetic stripe image. 9. The method according to claim 1, further comprising: before the step of acquiring the second image: Presenting prompt information on the terminal screen to prompt the subject to control the camera on the terminal to take pictures of the subject's face to obtain a second image when the subject can clearly see the first image; and/ Or prompting the detected person to hold the posture of the reference object; and/or prompting a way to adjust the distance between the terminal and the detected person. 10. An electronic device, characterized by comprising: at least one processor and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, so The instructions are executed by the at least one processor, so that the at least one processor executes the reference object-based diopter detection method according to any one of claims 1-9.